Extended documentation with MPS example.

docs/main.md

@@ -425,7 +425,7 @@ If the intel math kernel library should be integrated, one can switch on the par
 
 ## Cyclops tensor framework (ctf)
 
-[ctf](https://ctf.org) is a tensor framework for distributed parallelism. 
+[ctf](https://solomon2.web.engr.illinois.edu/ctf/) is a tensor framework for distributed parallelism. 
 This allows the computations to spread over several different compute cores and allows to use the hardware more efficient.
 It has an overhead for small tensors but allows to scale the compute power with very large tensors.
 

docs/snippets/s_mps.cpp

@@ -0,0 +1,33 @@
+#include "Xped/MPS/Mps.hpp"
+#include "Xped/MPS/MpsAlgebra.hpp"
+
+int main()
+{
+    // Alias for the symmetry
+    using Symmetry = Xped::Sym::SU2<Xped::Sym::SpinSU2>;
+
+    // Number of sites in the MPS
+    std::size_t L = 100;
+
+    // Amount of quantum number blocks
+    std::size_t Qinit = 10;
+
+    // Amount of basis states per quantum number block
+    std::size_t Minit = 100;
+
+    // Target quantum number of the MPS
+    Symmetry::qType Qtarget = Symmetry::qvacuum();
+
+    // local (physical) basis of the MPS. In this case one spin 1/2
+    Xped::Qbasis<Symmetry, /*depth=*/1> qloc;
+    qloc.push_back({2}, 1);
+
+    // Contruct the MPS (with random entries)
+    Xped::Mps<double, Symmetry> Psi(L, qloc, Qtot, Minit, Qinit);
+
+    // Compute the norm
+    auto norm = Xped::dot(Psi, Psi);
+
+    // Sweep from left to put all A-tensors into right-canonical form
+    for(std::size_t l = 0; l < L; ++l) { Psi.rightSweepStep(l, Xped::DMRG::BROOM::SVD); }
+}